Hard X-Ray Laser Reconstructs a Single Protein in 3D

Single-particle structural biology has long had an appealing XFEL promise: collect a diffraction pattern before an intense X-ray pulse destroys the molecule, then reconstruct structure from many randomly oriented particles. In practice, that workflow has remained far harder for proteins than for large viruses because a protein scatters weakly, aerosol delivery adds background and heterogeneity, and each exposure supplies only a single two-dimensional intensity measurement. The authors report a full single-particle imaging demonstration on erythrocruorin (Ery), a giant hemoglobin complex, using femtosecond hard-X-ray pulses.

Core Contribution

The central result is not merely a detection of isolated protein diffraction. The paper carries those patterns through classification and three-dimensional reciprocal-space merging to produce a 3D image at approximately 20 Å resolution. The reported reconstruction is coarser than atomic or near-atomic structural biology, but it crosses an important workflow boundary: a non-crystalline molecular complex is imaged as individual gas-phase particles rather than as a microcrystal or a large viral particle.

Ery is a useful intermediate target. Its double-doughnut architecture and approximately 170–270 Å dimensions make its low-resolution shape independently recognizable in diffraction features, while still placing the experiment much closer to protein-scale single-particle imaging than prior virus demonstrations. The paper therefore tests whether the combination of tighter focusing, aerosolized delivery, detector acquisition, and computational inference can retain enough signal for orientation-aware reconstruction.

Technical Approach

The experiment aerosolizes isolated Ery complexes and focuses them in a gas mixture selected to keep scattering background low. The particles intersect 6 keV XFEL pulses concentrated by Kirkpatrick–Baez mirrors into a 180 × 220 nm² spot. Each femtosecond exposure produces one diffraction pattern before radiation damage can evolve through the sample; the practical challenge is identifying the comparatively rare useful single-particle hits and combining them despite unknown particle orientations.

Figure 1 lays out this geometry, including the focused particle stream, the nanofocus beam, and the detector arrangement. The setup matters because signal collection is inseparable from delivery: background gas and variable particle states can obscure the already weak scattering from one protein complex.

The authors collected almost 200,000 diffraction patterns, then used expansion–maximization–compression (EMC) analysis to form 144 two-dimensional diffraction classes. The selected class averages show structured intensity rather than only radial scattering, providing the orientation-dependent information needed for a 3D merge. The central speckle in exemplary shots is consistent with Ery’s expected overall dimensions, which gives a direct sanity check before the more demanding reconstruction stage.

Figure 2 is the most persuasive intermediate evidence: the individual shots are noisy, but class averaging makes repeatable diffraction features visible across representative orientations. That is the bridge between single-hit detection and a volume reconstruction.

Results and Analysis

The paper reports 15 Å resolution for two-dimensional diffraction classes and approximately 20 Å for the resulting 3D image. Its merged three-dimensional reciprocal-space intensity extends beyond 20 Å, yet the real-space reconstruction does not realize all of that nominal reciprocal-space extent. The authors attribute this mismatch to heterogeneity introduced by gas-phase compaction of Ery complexes. That interpretation is plausible for a large, flexible complex delivered outside its native aqueous environment, but it also identifies the current bottleneck: collecting more signal alone will not fully recover resolution if the particles do not share one structure.

Figure 3 presents the merged Fourier intensity and the reconstructed 3D result. At this scale, the outcome supports recovery of global morphology rather than residue-level interpretation. The distinction is important. A roughly 20 Å image can validate particle shape and assembly-scale features, but it cannot yet compete with cryo-EM for detailed protein structure.

The evidence supports the narrower claim that hard-X-ray single-particle imaging can reconstruct a protein complex in three dimensions under these conditions. It does not yet establish a general route to native-state, high-resolution biomolecular imaging. Still, the experiment makes the constraints unusually concrete: throughput must rise, sample delivery must better preserve conformational state, and reconstruction must cope with particle-to-particle variation. If those elements improve together, XFEL imaging could offer time resolution unavailable to conventional cryogenic workflows and could complement mass-spectrometry studies with structural snapshots.